IMPLEMENTATION ANALYSIS OF ENCODING MANIPULATION METHOD IN DEFLATE COMPRESSION ALGORITHM

Abstract

In this paper, we use a novel Encoding Decision Chain Model to develop a theoretical method of encoding steganographic data in a Deflate compression algorithm. Specifically, we propose three major strategies that could be used for embedding covert data in a compressed data stream that rely on detecting decision points within the compression encoding process and manipulating their outcomes to achieve objectives secondary to the primary compression objective of reducing the amount of redundancy in source data as much as possible. The first technique, stored block length manipulation, artificially constraints the window of possible Deflate block lenghts to a smaller range and using those to embed secondary information. Second technique, back-reference manipulation, takes advantage of the ability to represent a repetitive string of symbols as either literals or back-references pointing at a specific chunk of previous data within the context window. By reducing the length of the back-reference, it is possible to emit several of them and utilize their lengths to encode steganographic information. The third technique, Huffman code tree structure manipulation, involves manipulating the values of relative frequencies of symbols of the alphabet stored within dynamic Huffman code compressed blocks to encode steganographic information. In this paper we include general analysis of each technique’s effectiveness with regards to the amount of encoded steganographic information against the amount of additional data needed to be stored within the carrier to ensure that no data is lost or altered as a result of proposed encoding manipulation techniques. Proposed methods have an advantage of not altering the carrier data directly, resulting in bit-perfect extraction of the original compressed data, reducing the surface for detecting this kind of steganographic apporach.